1 Introduction

Over the past two decades, several studies were developed regarding the presence of trace chemicals in wastewater treatment plant (WWTP) effluents and their harmful effects in human and environmental health (; Belhaj et al. 2015; Vandenberg et al. 2015). Some of those chemicals could act as endocrine-disrupting compounds (EDCs), interfering with the endocrine system by influencing the synthesis, release, transport, metabolism and excretion of hormones within the body of an affected organism (Fredj et al. 2015; Spina et al. 2015). Conventional WWTPs have low efficiencies for the removal of EDCs because they are designed to remove bulk constituents of wastewater, such as suspended solids, dissolved biodegradable organic matter, pathogens and nutrients, but are not equipped for dealing with trace pollutants (Ryu et al. 2011; Barber et al. 2015; Belhaj et al. 2015; Spina et al. 2015).

Numerous studies have surveyed the fate of EDCs in drinking water and WWTPs (Pauwels et al. 2008; Chang et al. 2009; Boles and Wells 2010; Ryu et al. 2011; Manickum and John 2014; Carvalho et al. 2015), claiming that WWTPs are a critical source in watercourses (e.g. rivers and surface waters) and important discharge points for EDCs. The behaviour of these pollutants along the different treatment stages within the WWTP is extremely important once the trace level concentration of EDCs creates a challenge for both the detection and removal processes, giving the studies an innovative character (Chang et al. 2009; Carvalho et al. 2015). Therefore, it is important to develop reliable wastewater treatment technologies that can efficiently remove these emerging contaminants at trace level concentrations, according to their different characteristics and their unpredictable behaviour in the different matrices under study.

The most common EDCs present in WWTP effluents are nonylphenol (NP) and octylphenol (OP), owing to the use of alkylphenol polyethoxylate non-ionic surfactants in industrial, commercial and domestic activities (Barber et al. 2015), as well natural estrogens, such as estrone, 17β-estradiol and ethinylestradiol (Diniz et al. 2010; Pessoa et al. 2014).

This work focuses on the presence of EDCs and their fate through a wastewater treatment plant, monitoring the presence of ten EDCs throughout the different stages of wastewater treatment. The aim of this work is to determine the fate of EDCs throughout a wastewater treatment plant process assessing the reliability of an analytical method already validated for monitoring a drinking water plant. The EDCs were analyzed in six different sampling points by solid-phase extraction (SPE) and ultra-performance liquid chromatography with electrospray ionization-tandem mass spectrometry (SPE-UPLC-ESI-MS/MS).

2 Material and Methods

2.1 Sampling

Monitoring and sampling was carried out during one month in the summertime of 2012: 20 samples were collected from a WWTP located in Lisbon, Portugal. This WWTP has primary, secondary and tertiary treatments, including filtration, as well as primary and secondary sedimentation and disinfection of the effluent for potential reuse and is able to treat 54,500 m3 per day of effluent, highlighting the large contribution from industrial effluents. Four samples were collected from each sampling point, in four different weeks. Composite samples were collected by an ISCO 6700 (Teledyne) autosampler every 24 h, at the same time, from the five chosen stages of the treatment and post-treatment effluent: (1) influent wastewater, (2) grit removal effluent, (3) primary sedimentation effluent, (4) secondary sedimentation effluent and (5) final effluent. A schematic diagram of the WWTP system showing the sampling sites is illustrated in Fig. 1.

Fig. 1
figure 1

Sampling points along the wastewater treatment plant: (1) influent; (2) grit removal; (3) primary sedimentation; (4) secondary sedimentation; (5) effluent

Full size image

2.2 Target Compounds

During this study ten EDCs were investigated including natural hormones (estriol, 17-β-estradiol and estrone), synthetic hormones (ethinylestradiol, progesterone, mestranol and diethylstilbestrol) and some industrial products (4-n-nonylphenol, 4-tert-octylphenol and bisphenol A), were analyzed in wastewater treatment plants samples. This trial was developed using ultra-performance liquid chromatography coupled to mass spectrometry and a triple quadrupole spectrometer (UPLC-ESI-MS/MS). Samples were previously pre-treated by vacuum filtration and solid-phase extraction (SPE).

2.3 Sample Preparation and Assay procedure

Sample preparation and analysis was carried out as previously described (Henriques et al. 2010; Carvalho et al. 2015). Wastewater samples were collected in 1000 mL glass amber bottles and transported to the laboratory under cooled conditions (4 °C). The samples were filtered and extracted as soon as they arrived in the laboratory, within 48 h after sampling, as described Henriques et al. (2010). Samples were hygiene cleaned by vacuum filtration with a 0.45-μm regenerated cellulose membrane to remove the suspended solids commonly found in this kind of effluent. Then, samples were prepared using a Dionex™ AutoTrace™ 280 (Thermo Scientific) automated solid-phase extraction system. Before liquid chromatographic analysis, extracts obtained by SPE were passed through an Acrodisc® syringe filter with a GHP membrane (13 mm, 0.45 μm) to remove some impurities that may compromise the analysis. Individual standards of each EDC (analytical grade <90 %) were obtained from Dr. Ehrenstorfer GmbH (Germany). A mixture of all EDC standards was prepared by combining the appropriate dilution of each individual stock solution and was subsequently used to prepare the working standard solutions for external calibration. The validated SPE-UPLC-ESI-MS/MS method (Henriques et al. 2010) was used to analyze the ten endocrine-disrupting compounds within each of the 24 WWTP samples.

2.4 Quality Assurance and Quality Control

The linearity of the method was tested using the calibration standards and by determining the limits of detection (LOD) and quantification (LOQ). LOD is the lowest analyte concentration expected to be reliably distinguished from the blank assay and at which detection is possible. The LOD is typically determined to be in the region where the signal to noise ratio is greater than 3. The assumption is that if analyte is present, it will produce a signal greater than the analytical noise in the absence of analyte. This is a simple and quick method. Similar interpretation is used to LOQ which indicates the level above which quantitative results may be obtained with a specified degree of confidence. This parameter is mathematically defined as equal to 10 times the standard deviation of the results for a series of replicates used to determine a justifiable LOD (Uhrovčík 2014).A correlation between the concentration value of each EDC and its correspondent peak area was acquired and linearization of the function allowed for the calculation of concentration values. Removal efficiencies were obtained according to Stasinakis et al. (2008), comparing concentration values before and after treatment.

2.5 2.5. Physico-Chemical Properties and Analytical Limits of Each Target Compound

This paper focuses on ten EDCs with different properties. Table 1 shows their physico-chemical properties, the detection and quantification limits and the main uses of each EDC.

Table 1 Physico-chemical properties and analytical limits of each target compound
Full size table

3 Results and Discussion

Analyses in quadruplicate of each sampling point were conducted to detect or quantify the target compounds under study along water treatment from the raw wastewater up to final effluent, in effluents from intermediary stages. The main results of each target compound’s concentrations obtained are presented in Table 2.

Table 2 Results obtained with SPE-UPLC-ESI-MS/MS analysis
Full size table

On examining the results obtained for the average concentration (ng/L) of each EDC in each sampling point, it is clearly visible that estrone (E1), 17-β-estradiol (E2), ethinylestradiol (EE2), diethylstilbestrol (DES), progesterone (P4) and octylphenol (OP) were not found in the 20 samples collected.

The average EDC concentration values for estriol (E3), nonylphenol (NP) and bisphenol A (BPA) detected throughout the wastewater treatment line are shown in Figs. 2, 3 and 4, respectively. At the first sampling point (influent), E3 could be detected and quantified, once LOD (3.30 ng/L) and LOQ (9.90 ng/L) were reached, and showed a range of concentrations between 3.09 and 11.07 ng/L (Fig. 2). After grit removal at sampling point 2, E3 concentrations decrease to 0.79–3.81 ng/L, achieving removal efficiency around 70 %. At sampling point 3, where the wastewater flows out of the primary sedimentation step, E3 concentration rose slightly (2.30–6.68 ng/L), overcoming the LOD value. However, when the effluent leaves the secondary sedimentation step (sampling point 4), concentrations decreased (0.14–0.36 ng/L), reaching values close to zero (0.04 ng/L) as the final WWTP effluent (sampling point 5), corresponding to a 98 % removal efficiency.

Fig. 2
figure 2

Occurrence of estriol (E3) in different sampling points

Full size image
Fig. 3
figure 3

Occurrence of nonylphenol (NP) in different sampling points

Full size image
Fig. 4
figure 4

Occurrence of bisphenol A (BPA) in different sampling points

Full size image

Nonylphenol (NP) was detected at similar concentrations at all sampling points of the treatment process, with concentration values above LOD (3.27 ng/L), which indicates a lack of removal of NP in the WWTP (Fig. 3). This is in accordance with literature, which suggests no removal, or even an increase, in NP from the liquid fraction of domestic and industrial mixed wastewater in the USA and China (Lian et al. 2009; Xia et al. 2011), having formed from degradation of NP ethoxylates. Results achieved could be justified by lower usage of this compound.

The fluctuation of MeEE2 concentrations between influents and effluents suggested potential internal sources (additives and return flows) or the influence of anomalies in the composite sampling. Nevertheless, none of the samples analyzed had concentration values (0.59–2.21 ng/L for influent and 1.56–4.10 ng/L for final effluent) higher than LOQ (10.40 ng/L) and, therefore, we will assume that MeEE2 was not quantified during this study. However, once LOD reaches 3.47 ng/L and the maximum concentration value reaches 4.10 ng/L, it is possible to confirm the presence of MeEE2 in the final effluent.

According with Fig. 4, bisphenol A (BPA) was efficiently removed during the treatment process. The concentration values of the raw WWTP influent ranged between 4.12 and 15.9 ng/L, BPA was detected and quantified once above LOD (3.27 ng/L) and LOQ (9.80 ng/L). After sampling point 2, BPA concentration decreased slightly, by 9 %. Overall, there is an exponential decrease on BPA concentrations, having removal efficiencies of 14, 20 and 56 %, corresponding to primary sedimentation, secondary sedimentation and final effluent. Compared to the raw wastewater, an overall removal efficiency of 67 % was achieved, confirming the results of prior European studies (Diniz et al. 2010; Tran et al. 2015). These results are very interesting, as previous studies have shown that even low BPA levels could considerably affect human and wildlife health (Vandenberg et al. 2015). In addition, our results are in accordance with studies conducted in other countries such as China, India, France, Brazil and Tunisia, where NP and BPA are the most abundant (Kumar et al. 2008; Qiang et al. 2013; Pessoa et al. 2014; Belhaj et al. 2015; Tran et al. 2015). Compared to the Annual Average-Environmental Quality Standards (AA-EQS) for nonylphenol and other priority substances (Directive 2013/39/EC), it is possible to conclude that wastewater treated in this study is in accordance with the legislated values for water policy, far below the maximum values allowed for discharge once the values obtained are expressed in ng/L, instead of μg/L. During the wastewater treatment process, removal of EDCs is considered to be a combination of biodegradation and adsorption, according to the physico-chemical characteristics of the target compound (Gaulke et al. 2008). For example, in regards to hydrophobicity, it is clear that the higher the octagonal-water partition coefficient (log Kow) the higher the hydrophobicity of a compound, which lends support to some of the conclusions reached by our study. However, considering the unpredictable characteristics of EDCs, it is important consider other aspects such as environmental conditions (temperature, pH, toxic compounds), the type of treatment procedure, sampling season and the occurrence of metabolites (Qiang et al. 2013; Pessoa et al. 2014; Belhaj et al. 2015; Tran et al. 2015;;), factors that have not been study during this research.

The results obtained are very encouraging, as both BPA and E3 were efficiently removed during the treatment process, allowing an understanding of the behaviour of EDCs throughout the wastewater treatment process, considering the composition of the raw water and fluctuations in concentration. Also, as Table 1 shows, BPA and E3 have lower log Kow, being the two compounds analyzed with lower hydrophobicity. This characteristic could justify higher removal efficiencies during WWTP treatments. This study is also one of the first to apply the UPLC-ESI-MS/MS methodology in a new matrix (wastewater); however, as this is still preliminary work, further analyses and validation are required to confirm the applicability of the method to be used in Portugal.

4 Conclusions

This research investigated the fate and analysis of ten endocrine-disrupting compounds, encompassing natural hormones, synthetic hormones and industrial products, in different sampling points throughout a wastewater treatment plant chain. The main conclusions obtained during this study emphasized that wastewater is an important contamination source of these compounds, the most common being BPA and nonylphenol NP. Most of the WWTPs in Portugal are not structured to remove EDCs. However, some simple procedures such as grit removal, primary sedimentation and secondary sedimentation could be useful in removing organic contaminants, such as E3 and BPA. It is important to note that this study was performed using an adapted version of a validated method developed for a drinking water matrix, not wastewater, so additional investigations are needed to support the preliminary outcomes obtained in this work, as well as further toxicological analysis. However, this research strengthens the importance of the precautionary principle and the importance of focusing WWTP treatments on the removal of endocrine-disrupting compounds. Nevertheless due to the unpredictable behaviour of these compounds, it is not easy to understand all the transformation processes and degradation rates occurring during treatment. Although, once NP, OP and other priority substances reach AA-EQS legislated levels of 0.30 μg/L (NP) and 0.10 μg/L, respectively, the concentration values obtained during this assays would not be considered an environmental warning, as we are dealing with smaller units, nanograms per litre instead of micrograms per litre.